Regardless of creation method, Fuzzy rules are of great importance in the implementation and optimization systems. Although using human knowledge in creating Fuzzy rules, has the advantage of readability and is near the experimental expertise, but it cannot be implemented in all systems. Since Output of a system is based on its correct function over the time, output data is reliable with higher percentage. In this paper, Fuzzy rules are extracted from a decision tree, constructed from the output of the system. In fact, traversing the decision tree leads to producing fuzzy rules. Decision tree which presented, is innovative, in comparison with previous implementations, and could also be regarded as new solution in classification. First advantage of the new decision tree to C4.5 (which is the most widely used as a common decision-making structure), is its capability of deciding on more than one feature simultaneously which is not provided in C4.5. Not producing a definite answer and result improvement in iterative processes are also other benefits of the new presented method.

Keywords